This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Durham University EXAMINATION PAPER date May/June 2008 exam code MATH2051/01 description Numerical Analysis II Time allowed: 3 hours Examination material provided: None Instructions: Credit will be given for the best FOUR answers from Section A and the best THREE answers from Section B. Questions in Section B carry TWICE as many marks as those in Section A. Approved electronic calculators may be used. ED01/2008 Durham University Copyright continued 2 MATH2051/01 SECTION A 1. The sequence { p n } generated by the formula p n +1 = 1 4 parenleftbigg 3 p n + 17 p 3 n parenrightbigg converges for suitable positive real values of p . To what limit does it converge and what is the order of convergence? 2. (i) A certain well behaved function f ( x ) has the following values . 40 . 44 . 48 . 52 . 56 . 1987 0 . 2182 0 . 2377 0 . 2571 0 . 2764 Form a divided difference table and compute f (0 . 462) as accurately as you can. (ii) Given that g ( x ) = 1 exp( x 2 ) 2 x sin x evaluate g (0 . 001) correct to three decimal places. 3. (a) Prove that for sufficiently smooth f , the error in the Central Difference formula takes the form f ( x + h ) f ( x h ) 2 h f ( x ) = h 2 6 f (3) ( c ) . (b) Using the approximation f ( x ) f ( x + h ) f ( x h ) 2 h and the data from Question 2(i) compute f (0 . 48) as accurately as you can. 4. Let A := parenleftbigg 1 10 4 10 4 1 parenrightbigg , b := parenleftbigg 10 4 10 4 parenrightbigg . Solve the linear system Ax = b with Gaussian elimination, once with and once without partial pivoting, using 3 digit arithmetic (meaning after each arithmetic operation the intermediate result is rounded to 3 significant figures)....
View Full
Document
 Fall '11
 Dr.A.Wacher
 Math, Numerical Analysis

Click to edit the document details